Merge pull request #47 from robusta-dev/helm-chart

Helm chart

.github/workflows/build.yaml

@@ -52,21 +52,7 @@ jobs:
           fixed-${{ github.sha }}
           fixed-
     - name: Build with skaffold
-      run: ./skaffold build -p prod --file-output=container-ids.json
-
-    # we disable the run-id so that if we do two identical builds then they will output the same YAML
-    # otherwise skaffold creates a different yaml because the run id changes
-    - name: Render yaml file
-      run: ./skaffold render -p prod --offline=true -l skaffold.dev/run-id=LEAVE_UNCHANGED > robusta.yaml
-
-    # upload robusta.yaml to gist
-    - name: Deploy gist
-      uses: arikalon1/actions-deploy-gist@v1
-      with:
-        token: ${{ secrets.GIST_TOKEN }}
-        gist_id: 6b809d508dfc3d8d92afc92c7bbbe88e
-        gist_file_name: robusta-${{ env.RELEASE_VER }}.yaml
-        file_path: ./robusta.yaml
+      run: ./skaffold build --file-output=container-ids.json
 
     - name: Create robusta examples
       run: |-

.github/workflows/test_robusta.yaml

@@ -17,6 +17,10 @@ jobs:
             source $HOME/.poetry/env
             cd src/
             poetry install
+        - name: Install Helm
+          run: |
+            curl https://raw.githubusercontent.com/helm/helm/master/scripts/get-helm-3 | bash
+            helm repo add robusta https://robusta-charts.storage.googleapis.com
         - name: Test Robusta
           env:
             PYTEST_SLACK_TOKEN: ${{ secrets.PYTEST_SLACK_TOKEN }}

src/Dockerfile → Dockerfile

@@ -9,23 +9,25 @@ RUN apt-get update \
 # disabled because of an issue with the libreadline7 dependency
 #RUN wget https://launchpad.net/~ionel-mc/+archive/ubuntu/socat/+build/15532886/+files/socat_1.7.3.2-2ubuntu2ionelmc2~ppa1_amd64.deb
 #RUN dpkg -i socat_1.7.3.2-2ubuntu2ionelmc2~ppa1_amd64.deb
-
-ENV CUSTOM_PLAYBOOKS_ROOT=/etc/robusta/config
 ENV ENV_TYPE=DEV
 
 # we install the project requirements and install the app in separate stages to optimize docker layer caching
 RUN mkdir /app
 RUN pip3 install --upgrade pip
 RUN pip3 install poetry==1.1.6
 RUN poetry config virtualenvs.create false
-COPY pyproject.toml /app
-COPY poetry.lock /app
+COPY src/pyproject.toml /app
+COPY src/poetry.lock /app
 WORKDIR /app
 RUN bash -c "pip3 install --requirement <(poetry export --dev --format requirements.txt --without-hashes)"
 
-COPY . /app
+ADD src/ /app
 
 RUN pip3 install --use-feature=in-tree-build .
 
+COPY playbooks/ /etc/robusta/playbooks/defaults
+RUN pip3 install -r /etc/robusta/playbooks/defaults/requirements.txt
+# remove the requirements so that we don't reinstall them at runtime
+RUN rm /etc/robusta/playbooks/defaults/requirements.txt
 # -u disables stdout buffering https://stackoverflow.com/questions/107705/disable-output-buffering
 CMD [ "python3", "-u", "-m", "robusta.runner.main"]

README.md

@@ -10,8 +10,10 @@ If you want to develop features for Robusta itself you'll need to install Robust
 
 If you want to use the Slack integration:
 1. Get a slack token from https://robusta.dev/integrations/slack/?id=xyz
-2. Add it to `deployment/dev/set_slack_api_key.yaml` before running skaffold
-3. Optional: reduce the chance you'll accidentally commit your Slack key to git by running `git update-index --skip-worktree deployment/dev/set_slack_api_key.yaml` (see [here](https://stackoverflow.com/a/40272289/495995) for details)
+2. Create your private `active_playbooks.yaml` file. 
+You can use `robusta gen-config --base-config-file helm/robusta/active_playbooks.yaml` to do that.
+3. Add the generated file content to `helm/robusta/active_playbooks.yaml` before running skaffold
+4. Optional: reduce the chance you'll accidentally commit your Slack key to git by running `git update-index --skip-worktree helm/robusta/active_playbooks.yaml` (see [here](https://stackoverflow.com/a/40272289/495995) for details)
 
 # Running Robusta runner locally as a developer
 If you want to run the Robusta runner on your own computer (e.g. with telepresence):

deployment/base/runner.yaml

@@ -1,73 +0,0 @@
-apiVersion: apps/v1
-kind: Deployment
-metadata:
-  name: robusta-runner
-  labels:
-    app: robusta-runner
-spec:
-  replicas: 1
-  selector:
-    matchLabels:
-      app: robusta-runner
-  template:
-    metadata:
-      labels:
-        app: robusta-runner
-    spec:
-      serviceAccountName: robusta-runner
-      containers:
-      - name: runner
-        image: us-central1-docker.pkg.dev/genuine-flight-317411/devel/robusta-runner
-        imagePullPolicy: Always
-        securityContext:
-          privileged: false
-        env:
-          - name: CUSTOM_PLAYBOOKS_ROOT
-            value: /etc/robusta/config
-          - name: ENABLE_MANHOLE
-            value: "true"
-          - name: ENABLE_COLORED_LOGS
-            value: "true"
-          - name: LOG_LEVEL
-            value: INFO
-          - name: INSTALLATION_NAMESPACE
-            valueFrom:
-              fieldRef:
-                fieldPath: metadata.namespace
-        volumeMounts:
-          - name: playbooks-config
-            mountPath: /etc/robusta/config
-        lifecycle:
-          preStop:
-            exec:
-              command: ["bash", "-c", "kill -SIGINT 1"]
-      - name: grafana-renderer
-        image: us-central1-docker.pkg.dev/genuine-flight-317411/devel/grafana-renderer:4
-        imagePullPolicy: Always
-        securityContext:
-          privileged: false
-        lifecycle:
-          preStop:
-            exec:
-              command: ["bash", "-c", "kill -SIGINT 1"]
-        resources:
-          limits:
-            memory: "512Mi"
-      volumes:
-        - name: playbooks-config
-          configMap:
-            name: robusta-config
-            optional: true
----
-apiVersion: v1
-kind: Service
-metadata:
-  name: robusta-runner
-spec:
-  selector:
-    app: robusta-runner
-  ports:
-    - name: http
-      protocol: TCP
-      port: 80
-      targetPort: 5000

docs/developer-guide/scheduled-playbooks.rst

@@ -39,6 +39,7 @@ Scheduled Playbook Example
         logging.info(f"My scheduled playbook running {action_params.some_string_param}")
 
 
+| Load the ``playbook`` code using ``robusta playbooks load ./playbooks``
 | Now, configure the playbook in the ``active_playbooks.yaml`` :
 
 .. code-block:: yaml
@@ -51,7 +52,7 @@ Scheduled Playbook Example
         action_params:
           some_string_param: "Scheduled Playbook - Hello World"
 
-| Lastly, deploy the playbook: ``robusta playbooks deploy ./playbooks``
+| Lastly, activate the playbook: ``robusta playbooks configure ./playbooks/active_playbooks.yaml``
 | That's it!
 | Now, this playbook will run every ``20`` seconds for ``10`` times.
 | In the robusta-runner logs, you'll be able to see the log line printed on each execution.

docs/getting-started/customization.rst

@@ -6,17 +6,19 @@ Lets see how we can customize the broad collection of builtin playbooks.
 
 Setting up a playbooks directory
 -------------------------------------------------------------
-You will need a `playbooks` directory with default files before continuing. There are three ways to set one up:
+Robusta comes with builtin `playbooks`
+Each ``playbook`` is a python script. Once loaded, it can be activated by including it in the ``active_playbooks.yaml`` file.
+The python defines the *available* playbooks and the yaml activates them.
 
-1. If you installed Robusta using ``robusta install``, you already have a `playbooks` directory in the working directory.
-2. You can download your `playbooks` directory from your cluster by running ``robusta playbooks pull``
-3. You can setup a new playbooks directory by running ``robusta examples``
+| To load a ``playbook`` script use ``robusta playbooks load my-custom-playbooks-dir``
+| To activate a ``playbook``, add it to your ``active_playbooks.yaml`` file, and run ``robusta playbooks configure active_playbooks.yaml``
+
+| You can download the builtin ``playbooks`` by running ``robusta examples``
+| To get the custom `playbooks` loaded to the cluster run ``robusta playbooks pull``
 
 .. warning::
     If you use ``robusta examples`` then `playbooks/active_playbooks.yaml` wont contain your API keys. Make sure to copy them over using ``robusta playbooks show-config``
 
-A playbooks directory contains python scripts and one file named ``active_playbooks.yaml``. The python defines the *available* playbooks
-and the yaml activates them.
 
 Enabling a new playbook
 ------------------------
@@ -44,7 +46,7 @@ From the playbooks directory run:
 
 .. code-block:: python
 
-   robusta playbooks deploy .
+   robusta playbooks configure active_playbooks.yaml
 
 The **deployment_babysitter** playbook is now running. You will get a notification in Slack every time a deployment's number of replicas changes.
 

docs/getting-started/installation.rst

@@ -16,16 +16,32 @@ Try running ``robusta --help`` to see what the robusta cli can do.
 
 Installing Robusta in a Kubernetes cluster.
 -----------------------------------------------------
-.. note:: ``robusta`` commands use your current kubectl context. Use ``kubectl config use-context`` to change it.
+.. note:: ``robusta`` and ``helm`` commands use your current kubectl context. Use ``kubectl config use-context`` to change it.
 
-To deploy Robusta to your cluster run:
+Robusta is installed using ``helm``.
+
+First, get a copy of Robusta's ``helm`` chart
+
+.. code-block:: bash
+
+    helm repo add robusta https://robusta-charts.storage.googleapis.com
+    helm pull robusta/robusta --untar
+
+Now, use the Robusta cli generate the initial configuration
+
+.. code-block:: bash
+
+    robusta gen-config --base-config-file robusta/active_playbooks.yaml
+
+Lastly, use that configuration to install robusta
 
 .. code-block:: bash
 
-   robusta install
+    cp active_playbooks_generated.yaml robusta/active_playbooks.yaml
+    helm install robusta ./robusta
 
-This will install two deployments in the ``robusta`` namespace. Robusta can be removed at any time by deleting that namespace. :ref:`Learn more about Robusta's architecture<Robusta Architecture>`.
-You can install Robusta on a custom namespace by using ``robusta install --namespace my-custom-namespace``
+This will install two deployments in the ``robusta`` namespace.
+Robusta can be removed at any time by running ``helm uninstall robusta``. :ref:`Learn more about Robusta's architecture<Robusta Architecture>`.
 
 Seeing Robusta in Action
 ------------------------------

docs/getting-started/more-commands.rst

@@ -3,6 +3,15 @@ More Robusta CLI Commands
 
 The ``robusta`` cli has additional commands for managing playbooks in your cluster.
 
+Initial configuration
+------------------------------------------
+To initially configure your ``active_playbooks.yaml`` file run
+
+.. code-block:: python
+
+   robusta gen-config --base-config-file robusta/active_playbooks.yaml
+
+
 Viewing active playbooks
 ------------------------------------------
 To view the currently active playbooks in your cluster run:

docs/getting-started/uninstallation.rst

@@ -1,5 +1,4 @@
 Uninstalling Robusta
 #####################
 
-Everything Robusta installs is placed in the ``robusta`` namespace. To delete Robusta, just delete that namespace.
-If you installed Robusta on a custom namespace, delete that one.
+Robusta is installed using ``helm``. Running ``helm uninstall robusta`` removes all the installed components.

docs/user-guide/architecture.rst

@@ -7,17 +7,15 @@ Robusta CLI
 -----------
 
 The robusta cli is installed via ``pip install robusta-cli`` and contains wrappers around kubectl to simplify
-robusta operations. For example, ``robusta install`` fetches Robusta's yaml manifests, customizes them with parameters
-from the user, and then runs ``kubectl apply``.
+robusta operations. For example, ``robusta playbooks configure`` loads the ``active_playbooks.yaml`` configuration into the cluster by creating a config-map.
 
 Robusta Kubernetes Deployments
 ------------------------------
 
 .. image:: ../images/arch.png
 
 
-All of Robusta's Kubernetes resources are installed in the ``robusta`` namespace.
-You can install robusta on a custom namespace, by using the ``namespace`` parameter on the ``robusta install`` command.
+All of Robusta's Kubernetes resources are installed in the ``robusta`` namespace by default.
 
 Robusta has two in-cluster Kubernetes deployments which trigger and execute playbooks.
 The first deployment, ``robusta-forwarder`` connects to the Kubernete's API server and monitors changes to the Kubernetes

docs/user-guide/playbook-configuration.rst

@@ -14,7 +14,8 @@ Here is the layout of an example playbook directory:
         - other_playbook.py
         - requirements.txt
 
-This set of playbooks would be loaded into Robusta with the command ``robusta playbooks deploy example_playbooks/``
+These playbooks code would be loaded into Robusta with the command ``robusta playbooks load example_playbooks/``
+This activation of these playbooks will be enabled using ``robusta playbooks configure example_playbooks/active_playbooks.yaml``
 
 At the moment, only one playbook directory can be loaded at a time. Loading another playbook directory will replace the previous one.
 

docs/user-guide/slack.rst

@@ -8,7 +8,7 @@ Recommended: Using Robusta's official Slack app
 Robusta is an approved app in the Slack App Directory. For details on Robusta's permissions,
 `see the Robusta page in the Slack App Directory <https://slack.com/apps/A0214S5PHB4-robusta?tab=more_info>`_
 
-To install the official Robusta app, use the ``robusta install`` command and follow the prompts.
+To install the official Robusta app, use the ``robusta gen-config`` command and follow the prompts.
 
 This method is recommended as it supports multiple Kubernetes clusters and are easy to setup.
 Please note that incoming Slack messages are routed through the official Robusta

helm/README.md

@@ -0,0 +1,24 @@
+# Robusta helm chart
+Robusta's helm chart is served from our gcp bucket
+
+# Upload a new chart version
+If you want to upload a new chart version, follow these steps:
+1. Update `Chart.yaml` file, with the updated `version` (Chart version) and `appVersion` (Robusta version)
+2. Make sure `helm` and `gsutil` is installed on your machine.
+3. From the directory, `helm`, run: `./upload_chart.sh`
+
+# Installing robusta with the helm chart
+1. Download the Robusta cli:
+`pip3 install -U robusta-cli --no-cache`
+2. Add Robusta's chart repo:
+`helm repo add robusta https://robusta-charts.storage.googleapis.com`
+3. Get a local copy of the Robusta chart
+`helm pull robusta/robusta --untar`
+This created a directory named `robusta` on your current directory
+4. Create the initial configuration for Robusta:
+`robusta gen-config --base-config-file robusta/active_playbooks.yaml`
+Follow the instructions, and a file named `active_playbooks_generated.yaml` is created
+5. Override the defaults file with the generated file:
+`cp active_playbooks_generated.yaml robusta/active_playbooks.yaml`
+6. Lastly, install Robusta:
+`helm install robusta ./robusta`